Unit 4

4.1.I2R Loss and Corona Loss

The average power-handling capacity of a 3-phase e.h.v. line and percentage lossdue to
I2R heating were discussed. Representative values are given below for comparison purposes.
System kV 400 750 1000 1150
Line Length, km 400 800 400 800 400 800 400 800
3-Phase MW/circuit When compared to the I2R heating loss, the average corona losses on
several lines from345 kV to 750 kV gave 1 to 20 kW/km in fair weather, the higher values
referring to highervoltages. In foul-weather, the losses can go up to 300 kW/km. Since, however,
rain does not fallall through the year (an average is 3 months of precipitation in any given
locality) andprecipitation does not cover the entire line length, the corona loss in kW/km cannot
be comparedto I2R loss directly. A reasonable estimate is the yearly average loss which amounts
to roughly2 kW/km to 10 kW/km for 400 km lines, and 20-40 kW/km for 800 km range since
usuallyhigher voltages are necessary for the longer lines. Therefore, cumulative annual average
coronaloss amounts only to 10% of I2R loss, on the assumption of continuous full load carried.
Withload factors of 60 to 70%, the corona loss will be a slightly higher percentage. Nonetheless,
during rainy months, the generating station has to supply the heavy corona loss and in some
cases it has been the experience that generating stations have been unable to supply full rated
load to the transmission line. Thus, corona loss is a very serious aspect to be considered in line
design640 320 2860 1430 6000 3000 8640 4320
(P = 0.5 V 2/xL)
% Power Loss = 50 r/x 4.98% 2.4% 0.8% 0.6%kW/km Loss, 3-phase 80 20 170 42.5 120 30 130
32.5.
When a line is energized and no corona is present, the current is a pure sine wave and capacitive.
It leads the voltage by 90°, as shown in Figure 4.1(a). However, when corona is present, it calls
for a loss component and a typical waveform of the total current is as shown in Figure 4.1 (b).
When the two components are separated, the resulting inphase component has a waveform which
is not purely sinusoidal, Figure 4.1 (c). It is still a current at power frequency, but only the
fundamental component of this distorted current can result in power loss.

Fig.4.1 Corona current waveform.

The mechanism of corona generation and its properties have been very extensively investigated
and the reader is referred to the bibliography at the end of the book. Of vital importance is the
generation of pulses which causes interference to radio, carrier communication, and gives rise to
TV interference. These aspects will be discussed in the next chapter. In this chapter, engineering
aspects of corona loss and audible noise will be described and data useful for design of lines
based on these two phenomena will be discussed.
4.2.Corona-Loss Formulae
2.1 List of Formulae
Corona-loss formulae were initiated by F.W. Peek Jr. in 1911 derived empirically from most
difficult and painstaking experimental work. Since then a horde of formulae have been derived
by others, both from experiments and theoretical analysis. They all yield the power loss as a
function of (a) the corona-inception voltage, Vo; (b) the actual voltage of conductor, V; (c) the
excess voltage (V — Vo) above Vo; (d) conductor surface voltage gradient, E; (e) corona
inception gradient, Eo; (f) frequency, f; (g) conductor size, d, and number of conductors in
bundle, N, as well as line configuration; (h) atmospheric condition, chiefly rate of rainfall, r,
and (i) conductor surface condition. The available formulae can be classified as follows: (see
Bewley and EHV Reference Books in Bibliography)
A. Those Based on Voltages
(i) Linear relationship : Skilling's formula (1931):

Pc μ V – Vo

(ii) Quadratic relationship

(a) Peek's formula (1911):
Pc μ (V – Vo)2

(b) Ryan and Henline formula (1924):

Pc μ V (V – Vo)

(c) Peterson's formula (1933) :

Pc μ V. F (V/Vo)

where F is an experimental factor.

(iii) Cubic Relationship

(a) Foust and Menger formula (1928):

Pc μ V3

(b) Prinz's formula (1940):

Pc μ V2(V – Vo)

B. Those Based on Voltage Gradients

(a) Nigol and Cassan formula (1961):
Pc μ E2 In (E/Eo)

(b) Project EHV formula (1966):

Pc μ V. Em, m = 5
In order to obtain corona-loss figures from e.h.v. conductor configurations, outdoor
experimental projects are established in countries where such lines will be strung. The resulting
measured values pertain to individual cases which depend on local climatic conditions existing
at the projects. It is therefore difficult to make a general statement concerning which formula or
loss figures fit coronal losses universally.We will here quote some formulas useful for
evaluating 3-phase corona loss in kW/km, which are particularly adopted for e.h.v. lines, and
some which are classic but cannot be used for e.h.v. lines since they apply only to single
conductors and not to bundles. There is no convincing evidence that the total corona loss of a
bundled conductor with N conductors is N times that of a single conductor.
(1) Nigol and Cassan Formula (Ontario Hydro, Canada).

Pc= K.f.r3. q .E2. ln (E/E0), kW/km, 3-ph ...

where f = frequency in Hz, r = conductor radius in cm.,

q = angular portion in radians of conductor surface where the voltage gradient exceeds the
critical corona-inception gradient,
E = effective surface gradient at operating voltage V, kV/cm, r.m.s.
E0= corona-inception gradient for given weather and conductor surface condition, kV/cm, r.m.s.
and
K = a constant which depends upon weather and conductor surface condition.
Many factors are not taken into account in this formula such as the number of sub-conductors in
bundle, etc.

(2) Anderson, Baretsky, McCarthy Formula (Project EHV, USA)

An equation for corona loss in rain giving the excess loss above the fair-weather loss in kW/3-
phase km is:

Here, PFW = total fair-weather loss in kW/km,

= 1 to 5 kW/km for 500 kV, and 3 to 20 kW/km for 700 kV,
K = 5.35 × 10–10 for 500 to 700 kV lines,
= 7.04 × 10–10 for 400 kV lines (based on Rheinau results),
V = conductor voltage in kV, line-line, r.m.s.,
E = surface voltage gradient on the underside of the conductor, kV/cm,peak,
r = rain rate in mm/hour,
r = conductor radius in cm, and N = number of conductors in bundle of each phase [The factor
0.3606 =1/1.609 3 ].
Formulae are not available for (a) snow, and (b) hoar frost which are typical of Canadian and
Russian latitudes. The EHV Project suggests K = 1.27 × 10–9 for snow, but this is a highly
variable weather condition ranging from heavy to light snow. Also, the conductor temperature
governs in a large measure the condition immediately local to it and it will be vastly different
from ground-level observation of snow.
The value of r to convert snowfall into equivalent rainfall rate is given as follows :
Heavy snow : r = 10% of snowfall rate;
Medium snow : r = 2.5% of snowfall rate;
Light snow : r = 0.5% of snowfall rate.
The chief disadvantage in using formulae based upon voltage gradients is the lack of difinition
by the authors of the formulae regarding the type of gradient to be used. there are several types
of voltage gradients on conductor surfaces in a bundle, such as nominal smooth-conductor
gradient present on a conductor of the same outer radius as the line conductor but with a smooth
surface, or the gradient with surface roughness taken into account, or the average gradient, or
the average maximum gradient, and so on. It is therefore evident that for Indian conditions, an
outdoor e.h.v. project is the only way of obtaining meaningful formulae or corona-loss figures
applicable to local conditions.
The Corona Current
The corona loss Pc is expressed as
Pc= 3 × line-to-ground voltage × in-phase component of current.
From the previously mentioned expressions for Pc, we observe that different investigators have
different formulas for the corona current. But in reality the current is generated by the
movement of charge carriers inside the envelope of partial discharge around the conductor. It
should therefore be very surprising to a discerning reader that the basic mechanism, being the
same all over the world, has not been unified into one formula for this phenomenon. We can
observe the expressions for current according to different investigators below.
1. Peek's Law. F.W. Peek, Jr., was the forerunner in setting an example for others to follow by
giving an empirical formula relating the loss in watts per unit length of conductor with nearly
all variables affecting the loss. For a conductor of radius r at a height H above ground,

where V, Vo are in kV, r.m.s., and r and H are in metres. The voltage gradients are, at an air
density of d, for a smooth conductor,

Attenuation of Travelling Waves Due to Corona Loss

A voltage wave incident on a transmission line at an initial point x = 0 will travel with a velocity
v such that at a later time t the voltage reaches a point x = vt from the point of incidence, as
shown in Figure 5.4 In so doing if the crest value of voltage is higher than the corona-disruptive
voltage for the conductor, it loses energy while it travels and its amplitude decreases
corresponding to the lower energy content. In addition to the attenuation or decrease in
amplitude, the waveshape also shows distortion. In this section, we will discuss only attenuation
since distortion must include complete equations of travelling waves caused by inductance and
capacitance as well as conductor and ground-return resistance.

Fig4.2 attenuation of voltage in transmission line

The energy of the wave is stored in both electromagnetic form and electrostatic form. The time
rate of loss of stored energy is equal to the power loss due to corona, whose functional
relationship with voltage has been given in Section 4.2. The total energy in a differential length
dx of the wave will be.
where L and C are inductance and capacitance per unit length of line. For a travelling wave, the
voltage V and current I are related by the surge impedance and the wave velocity is
Consequently,

The rate of dissipation of energy, assuming the capacitance does not change with voltage for the
present analysis, is

Now, the power loss over the differential length dx is

For different functional relations Pc= f(V), equation (5.29) can be solved and the magnitude of
voltage after a time of travel t (or distance x = vt) can be determined. We will illustrate the
procedure for a few typical values of f(V), but will consider the problem later on by using
equation (5.21) in Chapter 8.
(a) Linear Relationship
Let f(V) = Ks(V – V0).
Then, with Vi= initial voltage, 2 CV dV/dt = –Ks(V –V0). By separating variables and using the
initial condition V = Vi at t = 0 yields

where a = Ks/2C and V0= corona-inception voltage. Also, the voltages in excess of the corona-
inception voltage at any time t or distance x = vt will be

This expression yields an indirect method of determining a = Ks /2C by experiment, if distortion

is not too great. It requires measuring the incident wave magnitude Vi and the magnitude V after
a time lapse of t or distance x = vt at a different point on the line, whose corona-inception voltage
V0 is known. When C is also known, the constant Ks is calculated.
Audible Noise: Generation and Characteristics
When corona is present on the conductors, e.h.v. lines generate audible noise which is especially
high during foul weather. The noise is broadband, which extends from very low frequency to
about 20 kHz. Corona discharges generate positive and negative ions which are alternately
attracted and repelled by the periodic reversal of polarity of the ac excitation. Their movement
gives rise to sound-pressure waves at frequencies of twice the power frequency and its multiples,
in addition to the broadband spectrum which is the result of random motions of the ions, as
shown in Figure 5.5. The noise has a pure tone superimposed on the broadband noise. Due to
differences in ionic motion between ac and dc excitations, dc lines exhibit only a broadband
noise, and furthermore, unlike for ac lines, the noise generated from a dc line is nearly equal in
both fair and foul weather conditions. Since audible noise (AN) is man-made, it is measured in
the same manner as other types of man-made noise such as aircraft noise, automobile ignition
noise, transformer hum, etc.
 Fig 4.3 Audible noise frequency spectra from ac and dc transmission lines
Audible noise can become a serious problem from 'psycho-acoustics' point of view, leading to
insanity due to loss of sleep at night to inhabitants residing close to an e.h.v. line. This problem
came into focus in the 1960's with the energization of 500 kV lines in the USA.
Regulatory bodies have not as yet fixed limits to AN from power transmission lines since such
regulations do not exist for other man-made sources of noise. The problem is left as a social one
which has to be settled by public opinion. The proposed limits for AN are discussed in the next
section.
Limits for Audible Noise
Since no legislation exists setting limits for AN for man-made sources, power companies and
environmentalists have fixed limits from public-relations point of view which power companies
have accepted from a moral point of view. In doing so, like other kinds of interference, human
beings must be subjected to listening tests. Such objective tests are performed by every civic-
minded power utility organization. The first such series of tests performed from a 500-kV line of
the Bonneville Power Administration in the U.S.A. is known as Perry Criterion. The AN limits
are as follows:
No complaints : Less than 52.5 dB (A),
Few complaints : 52.5 dB (A) to 59 dB (A),
Many complaints : Greater than 59 dB (A),
The reference level for audible noise and the dB relation will be explained later. The notation (A)
denotes that the noise is measured on a meter on a filter designated as A-weighting network.
There are several such networks in a meter.
Design of line dimensions at e.h.v. levels is now governed more by the need to limit AN levels to
the above values. The audible noise generated by a line is a function of the following factors:
(a) the surface voltage gradient on conductors,
(b) the number of sub-conductors in the bundle,
(c) conductor diameter,
(d) atmospheric conditions, and
(e) the lateral distance (or aerial distance) from the line conductors to the point where noise is to
be evaluated.
The entire phenomenon is statistical in nature, as in all problems related to e.h.v. line designs,
because of atmospheric conditions.
While the Perry criterion is based upon actual listening experiences on test groups of human
beings, and guidelines are given for limits for AN from an e.h.v. line at the location of inhabited
places, other man-made sources of noise do not follow such limits. A second criterion for setting
limits and which evaluates the nuisance value from man-made sources of AN is called the 'Day-
Night Equivalent' level of noise. This is based not only upon the variation of AN with
atmospheric conditions but also with the hours of the day and night during a 24-hour period. The
reason is that a certain noise level which can be tolerated during the waking hours of the day,
when ambient noise is high, cannot be tolerated during sleeping hours of the night when little or
no ambient noises are present.
According to the Day-Night Criterion, a noise level of 55 dB(A) can be taken as the limit instead
of 52.5 dB(A) according to the Perry Criterion. From a statistical point of view, these levels are
considered to exist for 50% of the time during precipitation hours. These are designated as L50
levels.
AN Measurement and Meters
4.3.1 Decibel Values in AN and Addition of Sources
Audible noise is caused by changes in air pressure or other transmission medium so that it is
described by Sound Pressure Level (SPL). Alexander Graham Bell established the basic unit for
SPL as 20 × 10–6 Newton/m2 or 20 micro Pascals [2 × 10–5 micro bar]. All decibel values are
referred to this basic unit. In telephone work there is a flow of current in a set of head-phones or
receiver. Here the basic units are 1 milliwatt across 600 ohms yielding a voltage of 775 mV and
a current of 1.29 mA. For any other SPL, the decibel value is

This is also termed the 'Acoustic Power Level', denoted by PWL, or simply the audible noise
level, AN.
4.3.2 Microphones
Instruments for measurement of audible noise are very simple in construction in so far as their
principles are concerned. They would conform to standard specifications of each country, as for
example, ANSI, ISI or I.E.C., etc. The input end of the AN measuring system consists of a
microphone as shown in the block diagram, Figure 5.6. There are three types of microphones
used in AN measurement from e.h.v. lines and equipment. They are (i) air-condenser type; (ii)
ceramic; and (iii) electret microphones. Air condenser microphones are very stable and exhibit
highest frequency response. Ceramic ones are the most rugged of the three types.
Since AN is primarily a foul-weather phenomenon, adequate protection of microphone from
weather is necessary. In addition, the electret microphone requires a polarization voltage so that
a power supply (usually battery) will also be exposed to rain and must be protected suitably.

Fig 4.4; block diagram of AN measuring circuit

Some of the microphones used in AN measurement from e.h.v. lines are General Radio Type
1560-P, or 1971-9601, or Bruel and Kjor type 4145 or 4165, and so on. The GR type has a
weather protection. Since AN level form a transmission line is much lower than, say, aircraft or

ignition noise, 1" (2.54 cm) diameter microphones are used although some have used ones,
since these have more sensitivity than 1" microphones. Therefore, size is not the determining
factor.
The most important characteristic of a microphone is its frequency response. In making AN
measurements, it is evident that the angle between the microphone and the source is not always
90° so that the grazing angle determines the frequency response. Some typical characteristics
are as shown in Figure 5.7.

Fig 4.5; Response of microphone for grazing and perpendicular incidence

AN levels are statistical in nature and long-term measurements are carried out by protecting the
microphone from rain, wind, animals, and birds. Some types of shelters us wind-screens with a
coating of silicone grease. Foam rubber wind-covers have also been used which have negligible
attenuation effect on the sound, particularly on the A-weighted network which will be described
below. Every wind-cover must also be calibrated and manufacturers supply this data. Foam-
rubber soaks up rain and must be squeezed out periodically and silicone grease applied.
5.7.3 Weighting Networks
There are 5 weighting networks designated as A to E in Sound Pressure Level Meters. The 'A'
weighted network has been designed particularly to have nearly the same response as the human
ear, while the 'C' weighted network has a flat response up to 16 kHz. The 'A' network is also
least susceptible to wind gusts. It is also preferred by Labour Relations Departments for
assessing the adversity of noise-created psychological and physiological effects in noisy
environments such as factories, power stations, etc.
Typical frequency response of the A, B, C weighting networks are sketched in Figure 5.8(a)
while 5.8(b) compares the responses of A, D networks. The A-weighting network is widely
used for relatively nondirectional sources. From these curves it is seen that the C-network
provides essentially flat response from 20 Hz to 10 kHz. The human ear exhibits such flat
response for sound pressure levels up to 85 dB or more. At lower SPL, the human ear does not
have a flat response with frequency and the A and B networks are preferred. The A network is
used for SPL up to 40 dB and the B for SPL up to 70 dB. Sometimes, the A-weighted network
is known as the 40 dB network. It is also used for transformer noise measurements.

Fig 4.6; Frequency responses of (a) A,B,C weighting networks. (b) A,D weighting networks
5.7.4 Octave Band and Third Octave Band
It was mentioned earlier that in addition to the broadband noise generated by corona, pure tones
at double the power frequency and its multiples exist. These discrete-frequency components or
line spectra are measured on octave bands by selective filters.

Fig 4.7 Octave band AN meter circuit

The octave band consists of a centre frequency f0. Let f1 and f2 be the upper and lower
frequencies of the bands. Then An octave band extends from the lower frequenc
to the upper frequency A third-octave band extends from the lower
frequency to an upper frequency The
octave and third-octave band SPL is the integrated SPL of all the frequency components in the
band.
Corona Pulses: Their Generation and Properties
There are in general two types of corona discharge from transmission-line conductors: (i)
Pulseless or Glow Corona; (ii) Pulse Type or Streamer Corona. Both these give rise to energy
loss, but only the pulse-type of corona gives interference to radio broadcast in the range of 0.5
MHz to 1.6 MHz. In addition to corona generated on line conductors, there are spark
discharges from chipped or broken insulators and loose guy wires which interfere with TV
reception in the 80–200 MHz range. Audible noise has already been discussed in Chapter 5
which is caused by rain drops and high humidity conditions. Corona on conductors also causes
interference toCarrier Communication and Signalling in the frequency range 30 kHz to
500kHz.
In the case of Radio and TV interference the problem is one of locating the receivers far
enough from the line in a lateral direction such that noise generated by the line is low enough
at the receiver location in order to yield a satisfactory quality of reception. In the case of
carrier interference, the problem is one of determining the transmitter and receiver powers to
combat line-generated noise power. In this section we discuss the mechanism of generation and
salient characteristics of only pulse-type corona in so far as they affect radio reception. As in
most gas discharge phenomena under high impressed electric fields, free electrons and charged
particles (ions) are created in space which contain very few initial electrons. We can therefore
expect a build up of resulting current in the conductor from a zero value to a maximum or peak
caused by the avalanche mechanism and their motion towards the proper electrode. Once the
peak value is reached there is a fall in current because of lowering of electric field due to the
relatively heavy immobile space charge cloud which lowers the velocity of ions. We can
therefore expect pulses to be generated with short crest times and relatively longer fall times.
under dc excitation. Similar pulses occur during the positive and negative half-cycles under ac
excitation. The best equations that fit the observed wave shapes are also given on the figures. It
will be assumed that positve corona pulses have the equation

These equations have formed the basis for calculating the response of bandwidth-limited radio
receivers (noise meters), and for formulating mathematical models of the radio-noise problem.
In addition to the waveshape of a single pulse, their repetition rate in a train of pulses is also
important.

Fig 4.8 single positive and pulses

Referring to Fig. 4.8 when a conductor is positive with respect to ground, an electron avalanche moves
rapidly into the conductor leaving the heavy positive-ion charge cloud close to the conductor which
drifts away. The rapid movement of electrons and motion of positive-ions gives the steep front of the
pulse, while the further drift of the positive-ion cloud will form the tail of the pulse. It is clear that the
presence of positive charges near the positive conductor lowers the field to an extent that the induced
current in the conductor nearly vanishes. As soon as the positive-ions have drifted far enough due to
wind or neutralized by other agencies such as free electrons by recombination, the electric field in the
vicinity of the conductor regains sufficiently high value for pulse formation to repeat itself. Thus, a train
of pulses results from a point in corona on the conductor. The repetition rate of pulses is governed by
factors local to the conductor. It has been observed that only one pulse usually occurs during a positive
half cycle in fair weather and could increase to about 10 in rain where the water sprays resulting from
breaking raindrops under the applied field control electrical conditions local to the conductor.

Fig.4.10 formation of pulse train from positive polarity conductor

The situation when the conductor is negative with respect to ground is the reverse of that described
above. The electron avalanche moves away from the conductor while the positive-ion cloud moves
towards the negatively-charged conductor. However, since the heavy positive-ions are moving into
progressively higher electric fields, their motion is very rapid which gives rise to a much sharper pulse
than a positive pulse. Similarly, the lighter electrons move rapidly away from the conductor and the
electric field near the conductor regains its original value for the next pulse generation quicker than for
the positive case. Therefore, negative pulses are smaller in amplitude, have much smaller rise and fall
times but much higher repetition rates than positive pulses. It must at once be evident that all the
properties of positive and negative pulses are random in nature and can only be described through
random variables.

Pulses are larger as the diameter of conductor increases because the reduction in electric field strength
as one moves away from the conductor is not as steep as for a smaller conductor so that conditions for
longer pulse duration are more favourable. In very small wires, positive pulses can be absent and only a
glow corona can result, although negative pulses are present when they are known as Trichel Pulses
named after the first discoverer of the pulse-type discharge. Negative pulses are very rarely important
from the point of view of radio interference as will be described under "Radio Noise Meter Response" to
corona pulses.Therefore, only positive polarity pulses are important because of their larger amplitudes
even though their repetition rate is lower than negative pulses.
Measurement of RI, RIV, and Excitation Function
The intereference to AM broadcast in the frequency range 0.5 MHz to 1.6 MHz is measured in
terms of the three quantities : Radio Interference Field Intensity (RIFI or RI), the Radio Influence
Voltage (RIV), and more recently through the Excitation Function. Their units are mV/m, mV,
and mA/ m or the decibel values above their reference values of 1 unit ( V / m, V,mmm A / m ).
The nuisance value for radio reception is governed by a quantity or level which is nearly equal to
the peak value of the quantity and termed the Quasi Peak. A block diagram of a radio noise
meter is shown in Fig. 4.11. The input to the meter is at radio frequency (r-f) which is amplified
and fed to a mixer. The rest of the circuit works exactly the same as a highly sensitive super-
heterodyne radio receiver, However, at the IF output stage, a filter with 5 kHz or 9 kH bandwidth
is present whose output is detected by the diode D. Its output charges a capacitance C through a
low resistance Rc such that the charging time constant Tc= Rc C = 1 ms. A second resistance Rd
is in parallel with C which is arranged to give a time constant Td= Rd C = 600 ms in ANSI
meters and 160 ms in CISPR or European standard meters. Field tests have shown that there is
not considerable difference in the output when comparing both time constants for line-generated
corona noise. The voltage across the capacitor can either be read as a current through the
discharge resistor Rd or a micro-voltmeter connected across it.

Fig.4.11 Block diagram of Radio noise meter

For radiated interference measurement RI, the front end of the meter is fitted with either a rod
antenna of 0.5 to 2 metres in length or a loop antenna of this size of side. For conducted
measurements, the interfering voltage RIV is fed through a jack. The input impedance of the
meter is 50 ohms:
The following formulas due to Nigol apply to the various settings of the noise meter for
repetitive pulses:

The object under test, which could be an insulator string with guard rings, is energized by a high
voltage source at power frequency or impulse. A filter is interposed such that any r-f energy
produced by partial discharge in the test object is prevented from flowing into the source and all
r-f energy goes to the measuring circuit. This consists of a discharge-free h.v. coupling capacitor
of about 500 to 2000 pF in series at the ground level with a small inductance L. At 50 Hz, the
coupling capacitor has a reactance of 6.36 Megohms to 1.59 Megohms. The value of L is chosen
such that the voltage drop is not more than 5 volts so that the measuring equipment does not
experience a high power-frequency voltage.
Let V = applied power frequency voltage from line to ground,
VL = voltage across L,
Xc= reactance of coupling capacitor and XL = 2pfL = reactance of inductor.

Fig 4.12 circuit for measuring radio influence voltage(riv)

At radio frequencies, the inductance presents a very high impedance while the coupling capacitor
has very low reactance. The capacitor is tuned at a fixed frequency, usually 1 MHz, with an r-f
choke, L0. There is a series Rg to ground. This r-f voltage is fed to the noise meter through a
length of cable of 50 ohm characteristic impedance terminated in a 50 ohm resistance at the input
end of the meter.
The r-f voltage developed across Rg is fed to the noise meter. Since transmission lines have a
characteristic impedance in the range 300 to 600 ohms, standard specifications stipulated that the
r-f voltage must be measured across 600 ohms. Thus, Rg is in the neighbourhood of 600 ohms.
However, it was obvious that this could not be done since the presence of cable will lower the
impedance to ground. In earlier days of RIV measurement at lower voltages (230 kV equipment)
the noise meter was directly connected across Rg and its input end was open. The operator sat
right underneath the pedestal supporting the coupling capacitor in order to read the meter or used
a pair of binoculars from a distance. But with increase in test voltage, the need for maintaining a
safe distance necessitates a cable of 10m to 20m. With its surge impedance Rc conected across
Rg, which has a higher value, the combined parallel impedance is lower than Rc. No coaxial
cables are manufactured for high surge impedance, so that standard specifications allow RIV to
be measured across 150 ohm resistance made up of Rg and Rc in parallel. It is clear that the
measuring cable must have Rc greater than 150 ohms. The highest impedance cable has Rc= 175
ohms on the market. The value of Rg can be selected such that